Security AffairsPierluigi Paganini
Security Affairs rounds up AI cybersecurity developments
Security Affairs published its seventh AI-cybersecurity newsletter on 8 November 2026, collecting links to research and news about the use of artificial intelligence in cyber attacks and defence. It says AI agents can automate tasks, analyse large volumes of data, find vulnerabilities and speed up offensive operations, while also supporting threat detection, incident analysis and response. The newsletter is presented as a weekly selection of developments at the intersection of AI and cybersecurity.
The linked items cover a range of subjects, including a report that an Accenture contractor connected an AI agent to the US Pentagon’s secure network and sent data to a high-risk website; Cursor’s security notice about malicious versions 2.1.312 and 2.1.313 of Anthropic[.]claude-code; and Cisco Talos research on AI-based victim prioritisation by the VOLTA infostealer.
Other selections address LLMjacking, prompt adware concealed in jailbreak bundles, a WorkOS data security incident dated 30 October 2026, and a flaw in Claude Dispatch that researchers say could enable browser extensions to execute code on the operating system. The newsletter also links to material on military AI cooperation between the UK and Germany, AI safety research, and security measures for AI developer platforms.